Teardown

Daily digest · 2026-08-25

Scan #038: The middlemen taking rebates, and the wedges pricing the middle out

Four incumbents whose margins depend on being between a manufacturer and a buyer — and four emerging companies trying to reprice the layer they sit on.

Today’s eight all sit in the middle of a value chain and get paid for coordination. The four incumbents — a used-car retailer, an electrical distributor, a paint chain, and a regulated utility — earn a spread between a supplier and a buyer, and each is being asked whether the spread is a moat or a subsidy. The four emerging companies — a construction payments layer, a cafe supply platform, an embedded insurance API, and a rent-payment loyalty program — are each trying to reprice a specific piece of that middle before an incumbent notices the volume or a partner stops writing the check.

CarMax — Retail · Incumbent, at risk. The largest US used-car retailer, ~256 superstores, ~$26B FY2026 revenue and a $16.4B captive-finance book. But retail comp used units went negative in Q1 FY2027, Carvana just posted 596,641 units at +43% growth on $20.3B, and CAF loss provisions on 2022-23 vintages keep grinding higher. A brand-new outsider CEO (Keith Barr, ex-IHG, started March 2026) is running a four-pillar reset while Carvana takes the share.

Wesco International — Logistics / Distribution · Incumbent, at risk. The 1922 Westinghouse spin-out that PE flipped twice, went public in 1999, and used a 2020 Anixter merger to become a $23.5B electrical/data-comm/utility distributor. Data-center backlog is camouflaging the rest of the story: soft organic growth in EES and UBS, rebate-dependent gross margins that Amazon Business and procurement platforms can strand, and post-Anixter net leverage still near 3x adjusted EBITDA. A middleman with a good year, not a widening moat.

Sherwin-Williams — Construction / Paint & Coatings · Incumbent, well positioned. The 1866 Cleveland paint pioneer is the only major architectural coatings player that OWNS its distribution — ~4,900 company-operated Pro-first stores in North America, 46 straight years of dividend hikes, 14 of the last 15 quarters of gross-margin expansion. PPG has been cutting price and Home Depot has pushed Behr Pro, and share still moves Sherwin’s way because the store network is the moat.

Xcel Energy — Energy / Regulated Utility · Incumbent, well positioned. A four-state regulated utility (~3.9M electric, ~2.1M gas) sitting on one of the cleanest AI-power rate-base ramps in the sector: ~$56B → ~$94B rate base at 11% CAGR through 2030, powered by Google’s 1,900 MW Minnesota contract and 950 MW of new Colorado data-center generation. The $640M Marshall Fire settlement ($350M insurance-recovered) and the Texas AG’s Smokehouse Creek suit are real tails — but the regulated-recovery mechanism is what actually decides it.

Handle — Construction / Fintech · Emerging. A construction back-office platform for preliminary notices, lien waivers, and integrated payments, sold to the credit teams at Ferguson, Home Depot/SRS/GMS, Cemex, ABC Supply and Heidelberg Materials. Compliance is the wedge; the question is whether Handle can escape being sold as a stand-alone Levelset alternative and convert the credit-team footprint into a real payments rail before Procore Pay’s default GC distribution and Trimble/Flashtract close the loop from the other side.

Odeko — Retail / SMB Wholesale · Emerging. An overnight-delivery, single-portal supply platform for ~16,000 independent coffee shops — reportedly ~21% COGS savings for the operators. Real cost line, thin attach base. The open question is whether the density Odeko has proved on cafes is defensible before Sysco, US Foods CHEF’STORE or Restaurant Depot notice it and price the same SKUs 20-30% cheaper into the same accounts.

Cover Genius — Insurance / Embedded · Emerging. Sydney-founded embedded insurance platform with the deepest partner distribution stack in the category: a single XCover API, licences in 60+ countries and all 50 US states, and integrations inside Booking.com, Uber, Klarna, Ryanair and eBay. Latest mark ~$1.9B on Vista Credit Partners debt. Whether the licence-and-integration footprint is an actual moat depends on whether Amazon/Uber/Booking eventually insource the layer and whether loss ratios can hold at the take-rate partners demand.

Bilt Rewards — Retail / Consumer Loyalty & Payments · Emerging. The rent-payment loyalty program that convinced Wells Fargo to underwrite roughly $120M/yr of card losses to sustain a $10.75B mark — and now has to reprice the whole model as Wells steps back and Cardless steps in. The wedge (rent as the anchor spend) is genuinely novel; whether landlord, merchant and issuer economics all pay at once is the actual question, and the Wells write-down is the thing that concentrates the mind.

Full deep dives